DE102013209093A1 - Method for producing a mask for a lithographic illumination system - Google Patents

Abstract

Method for producing a mask for a lithographic illumination system, comprising the following steps: a) setting an intensity variation device in such a way that an illumination intensity that is as uniform as possible is provided; b) determining an illumination direction distribution at a defined number of field points (PM) on the mask as a function of a first coordinate after averaging over a second coordinate; c) calculating structure sizes at the field points (PM), so that structures of the desired size are obtained in the resist image; and e) creating the mask, the structure size (b) on the mask at one location being given by a defined interpolation of structure sizes (b) on the mask at the field points (PM).

Description

The invention relates to a method for producing a mask for a lithographic illumination system.

State of the art

The use of prior art masks for lithographic lighting systems results in high frequency errors of the structures printed on a wafer due to inhomogeneities of the illumination of the mask in the illumination system used. In this context, "high-frequency" means a higher spatial frequency than corresponds to the possible resolution of an intensity variation device present in the illumination system. At a certain number of field points, the structures are printed with the desired widths, but deviations of the printed structure widths to the desired widths occur between these field points.

Disclosure of the invention

It is an object of the present invention to provide an improved method of making masks for a lithographic lighting system.

• a) Einstellen einer Intensitätsvariationseinrichtung derart, dass eine möglichst gleichmäßige Beleuchtungsintensität bereitgestellt wird;
• b) Bestimmen einer Beleuchtungsrichtungsverteilung an einer definierten Anzahl von Feldpunkten auf der Maske als Funktion einer ersten Koordinate nach Mittelung über eine zweite Koordinate;
• c) Berechnen von Strukturgrößen an den Feldpunkten, so dass im Lackbild Strukturen der gewünschten Größe erhalten werden; und
• d) Erstellen der Maske, wobei die Strukturgröße auf der Maske an einem Ort gegeben ist durch eine definierte Interpolation von Strukturgrößen auf der Maske an den Feldpunkten.

The object is achieved with a method for producing a mask for a lithographic illumination system, comprising the following steps:

• a) setting an intensity variation device such that the most uniform possible illumination intensity is provided;
• b) determining a lighting direction distribution at a defined number of field points on the mask as a function of a first coordinate after averaging over a second coordinate;
• c) calculating structure sizes at the field points, so that structures of the desired size are obtained in the paint image; and
• d) Creating the mask, wherein the feature size on the mask is given at a location by a defined interpolation of feature sizes on the mask at the field points.

• a) Einstellen einer Intensitätsvariationseinrichtung derart, dass eine möglichst gleichmäßige Beleuchtungsintensität bereitgestellt wird;
• b) Bestimmen einer Beleuchtungsrichtungsverteilung an einer definierten Anzahl von Feldpunkten auf der Maske als Funktion einer ersten Koordinate nach Mittelung über eine zweite Koordinate;
• c) Berechnen von Strukturgrößen an den Feldpunkten, so dass im Lackbild Strukturen der gewünschten Größe erhalten werden;
• d) Bestimmen der Beleuchtungsintensität an einer möglichst großen Anzahl von Feldpunkten; und
• e) Erstellen der Maske unter Berücksichtigung einer Korrektur der Strukturgrößen gemäß folgender mathematischer Beziehung:

mit den Parametern:

Δb

Korrektur der Strukturgröße

CD

gewünschte Strukturbreite

MEEF

Mask-Error-Enhancement-Factor

NILS

Normalised-Intensity-Logarithm-Squared

β

Vergrößerungsmaßstab

ΔI(x)/I

Restintensitätsfehler als Funktion der x-Koordinate

According to a second aspect, the object is achieved by means of a method for producing a mask for a lithographic illumination system, comprising the following steps:

• a) setting an intensity variation device such that the most uniform possible illumination intensity is provided;
• b) determining a lighting direction distribution at a defined number of field points on the mask as a function of a first coordinate after averaging over a second coordinate;
• c) calculating structure sizes at the field points, so that structures of the desired size are obtained in the paint image;
• d) determining the illumination intensity at the largest possible number of field points; and
• e) Creation of the mask taking into account a correction of the feature sizes according to the following mathematical relationship:

with the parameters:

.DELTA.b

Correction of the structure size

CD

desired structure width

MEEF

Mask error enhancement factor

NILS

Normalized Intensity Logarithm-Squared

β

magnification

.DELTA.I (x) / I

Residual intensity error as a function of the x-coordinate

• a) Einstellen einer Intensitätsvariationseinrichtung derart, dass eine möglichst gleichmäßige Beleuchtungsintensität bereitgestellt wird;
• b) Bestimmen einer Beleuchtungsrichtungsverteilung an einer definierten Anzahl von Feldpunkten auf der Maske als Funktion einer ersten Koordinate nach Mittelung über eine zweite Koordinate;
• c) Berechnen von Strukturgrößen an den Feldpunkten, so dass im Lackbild Strukturen der gewünschten Größe erhalten werden;
• d) Bestimmen der Beleuchtungsintensität an einer möglichst großen Anzahl von Feldpunkten;
• e) definiertes Interpolieren von Strukturgrößen auf der Maske an den jedem Feldpunkt nächstgelegenen beiden Feldpunkten; und
• f) Erstellen der Maske unter Berücksichtigung einer Korrektur der Strukturgrößen gemäß folgender mathematischer Beziehung:

mit den Parametern:

Δb

Korrektur der Strukturgröße auf der Maske

CD

gewünschte Strukturbreite

MEEF

Mask-Error-Enhancement-Factor

NILS

Normalised-Intensity-Logarithm-Squared

β

Vergrößerungsmaßstab

ΔI(x)/I

Restintensitätsfehler als Funktion der x-Koordinate

According to a third aspect, the object is achieved by means of a method for producing a mask for a lithographic illumination system, comprising:

• a) setting an intensity variation device such that the most uniform possible illumination intensity is provided;
• b) determining a lighting direction distribution at a defined number of field points on the mask as a function of a first coordinate after averaging over a second coordinate;
• c) calculating structure sizes at the field points, so that structures of the desired size are obtained in the paint image;
• d) determining the illumination intensity at the largest possible number of field points;
• e) defined interpolation of feature sizes on the mask at the two field points closest to each field point; and
• f) Creation of the mask taking into account a correction of the feature sizes according to the following mathematical relationship:

with the parameters:

.DELTA.b

Correction of the structure size on the mask

CD

desired structure width

MEEF

Mask error enhancement factor

NILS

Normalized Intensity Logarithm-Squared

β

magnification

.DELTA.I (x) / I

Residual intensity error as a function of the x-coordinate

Preferred embodiments of the method according to the invention are the subject of subclaims.

An advantageous development of the method according to the invention provides that in step b) the determination of the illumination direction distribution is carried out at all field points at which an independent intensity adaptation is possible. Advantageously, in this way according to the invention, interpolations are carried out between a large number of field points and thus even small CD errors are largely compensated.

A further preferred embodiment of the method according to the invention provides that in step b) the determination of the illumination direction distribution at a number of field points (PM) in the range between three and five is performed. With this specific selection of field points, a reduction of mask design field points is performed, thereby significantly reducing the computational effort.

A further advantageous development of the method according to the invention provides that in step e) between the field points (PM) the interpolation of the feature sizes is one of: linear, square, parabolic. Advantageously, several known interpolation methods can thereby be carried out, whereby very variable interpolation requirements can be taken into account.

A further advantageous development of the method according to the invention provides that in step b) the determination of the illumination direction distribution is carried out at one or at all field points at which an independent intensity adjustment is possible. Advantageously, different levels of improvement of the mask correction can be lifted.

Further preferred embodiments of the method according to the invention provide that the field points (P) are arranged in the center of the mask +/- of the 2/3 reticle radius or one of the field points at the edge of the mask. In this way, it can be advantageously taken into account in which region of the mask the correction requirement for the mask is greatest.

Advantages of the invention

The invention will be described in detail below with further features and advantages with reference to several figures. All described or illustrated features alone or in any combination form the subject of the invention, regardless of their summary in the claims or their dependency, and regardless of their formulation or representation in the Description or in the figures. The figures are primarily intended to illustrate the principles essential to the invention.

In the figures shows:

1a and 1b CD error in a mask design at each field point PU, at which an independent intensity adjustment is possible, or only in the middle of the field (P0), evaluated at all field points PU, where an independent intensity adjustment is possible

2a and 2 B CD error in a mask design at each field point PU, at which an independent intensity adjustment is possible, or only in the middle of the field (P0), evaluated at all field points;

3 an illumination intensity profile over the mask after correction by means of an intensity variation device;

4a to 4c CD error in a conventional mask design only in the middle of the field;

5a to 5c according to the invention reduced CD errors in a linear interpolation of feature sizes in a mask design in the field center and +/- 2/3 field width;

6a to 6c according to the invention reduced CD errors in a linear interpolation of feature sizes in a mask design in the middle of the field and at the edge of the field;

7a to 7c according to the invention reduced CD errors in a quadratic interpolation of feature sizes in a mask design in the middle of the field and +/- 4/5 field width;

8a to 8c CD error in a conventional mask design at each position of fingers of the intensity variation device;

9a to 9c According to the invention reduced CD errors in a correction of the feature widths in a mask design only in the middle of the field;

10a to 10c according to the invention, reduced CD errors with a correction according to equation (5) and subsequent linear interpolation with a mask design in the middle of the field and +/- 2/3 field width;

11a to 11c according to the invention, reduced CD errors with a correction according to equation (5) and subsequent linear interpolation with a mask design in the middle of the field and at the edge of the field;

12a to 12c according to the invention, reduced CD errors with a correction according to equation (5) and subsequent quadratic interpolation with a mask design in the middle of the field and +/- 4/5 field width; and

13a to 13c Reduced CD errors according to the invention in a correction according to equation (5) in the case of a mask design on each finger of the intensity variation device.

Embodiments of the invention

Lithographic illumination systems typically include an intensity variation device having a plurality of, for example, 25 fingers, each of the fingers having a defined width that is equal to the value of the distance that the individual fingers of the intensity varying device are from each other. Each finger has a curvature corresponding to a projection of the curvature of the object field of the projection lens at the front edge, ie where the finger is pushed into the illumination light. It is not taken into account that the fingers partly overlap due to geometric conditions.

In the following, far fields of three different light sources A, B and C with different light characteristics (eg plasma sources with different geometrical configurations or orientations) are considered.

It is known that a mask must first be designed for printing desired structures in a lithographic process. This means that feature sizes are calculated on the mask based on the knowledge of the desired structures to be printed on the wafer as well as an assumption about illumination intensity and illumination direction distribution ("illumination pupil"). If the mask is illuminated precisely with this illumination pupil and illumination intensity during the lithography process, structures result on the wafer which have exactly the desired size. Frequently, the assumed illumination intensity and / or illumination direction distribution differs from the actual illumination intensity and / or illumination direction distribution, because considering the actual illumination intensity and / or illumination direction distribution would be too complicated.

The structures considered in the following embodiment are 18 nm lines with a pitch between 36 nm and 126 nm. 100 different structures are considered, in each case 50 different period lengths in horizontal and vertical orientation.

Before the mask design, a locally substantially constant intensity profile of the illumination of the reticle is adjusted by means of the intensity variation device, so that a uniformity is corrected as best as possible. This can also be done by simulation. In this case, a position of the fingers of the intensity variation device is sought in such a way that substantially the same scan-integrated intensity is obtained at all field points P. Since the fingers can only be moved along one direction and thus only one degree of freedom is present, only the intensity at a field point PU can be freely selected per finger. It is assumed that this field point PU is in each case in the middle of the finger. Of all field points P, a desired intensity can be ensured only at certain field points PU whose number is determined by the number of fingers.

• 1. An jedem der beispielsweise 25 Feldpunkte PU wird die ideale Maskenauslegung bestimmt, d. h., dass pro Finger der Intensitätsvariationseinrichtung eine eigene Maskenauslegung durchgeführt wird. Die entsprechende Maske wird an jedem Feldort P verwendet, der vom betreffenden Finger überdeckt wird.
• 2. Nur in Feldmitte (Feldpunkt P0) wird eine Maskenauslegung durchgeführt, wobei diese Maske dann für jeden Feldort verwendet wird.

Since the illumination direction distribution and the illumination intensity are not locally constant, a mask layout under the conditions given above would result in a different pattern width on the mask at each field location if the same structure is to be printed on the wafer. The following are examined, known in the prior art two options:

• 1. At each of the example 25 field points PU, the ideal mask design is determined, ie, a separate mask design is performed per finger of the intensity variation device. The corresponding mask is used at each field location P, which is covered by the respective finger.
• 2. Only in the middle of the field (field point P0) is a mask design carried out, this mask then being used for each field location.

The 1a and 1b show qualitatively CD errors (ie feature size deviations) as a function of the reticle coordinate R. This coordinate is orthogonal to a direction of displacement of the mask during scanning.

One recognizes in 1a in that, in the event that a mask design is carried out at each field point PU, the desired structure widths are produced at each field point PU for all the structures on the wafer that are considered. The CD error is thus equal to zero at all field points PU.

Will, however, as in 1b shown, only at a single field point P0, namely performed in the middle of the field (at R = 0 mm) a mask design, so the CD error for all structures is only zero in the middle of the field. For all other field points a finite CD error results. For each field point PU, 100 points are drawn in (corresponding to one point per structure), which however partially overlap in the figure and therefore not all are recognizable as individual points.

The CD error for field points P that lie between the field points PU is in 1 not shown.

The mask design is based on a calculation of aerial images and is therefore numerically demanding. Therefore, mask manufacturers use rule-based mask adjustments to reduce computational effort. In these, a mask designed by means of an aerial image calculation is taken and adapted by means of simple numerical rules. Due to the simplicity of these rules, this type of mask adaptation advantageously requires only a small amount of computation time.

The CD error for field points P lying between the field points PU was in 1 not shown. In the 2a and 2 B are additionally qualitatively CD errors for field points P, which lie between the field points PU. The curves show the CD error for the different structures as a function of the field location. There are 100 curves drawn, each individual line corresponding to a structure. Of the Value of the CD error at the field points PU is marked by dots, which means that each point on the corresponding field point PU lies on the corresponding curve.

It can be seen that even with a mask design at each field point PU (see 2a ) gives a significant CD error. The main cause for this is the limitability of uniformity of the local intensity that can be achieved by the intensity variation device. The CD error is even more pronounced in the middle of the field in the case of a mask design (see 2 B ).

Deviations of the illumination intensity after an above-described correction of the intensity profile has been performed by means of an intensity varying device, are distributed over the reticle in 3 presented qualitatively. At the field points P between the centers PU of the fingers, the corresponding finger has a position that is not optimal. The jumps in the curve mark the locations where a finger of the intensity varying device ends and an adjacent finger begins.

In a systematic investigation of the CD errors achieved by means of conventional mask designs, the errors in the CD errors result in the CD errors 4a to 4c and 8a to 8c qualitatively presented results.

Shown in the 4a to 4c and 8a to 8c three gradients of CD errors for three different light sources A, B and C with different light characteristics. The different light sources lead to different intensity profiles and illumination direction distributions, with which the mask is illuminated.

In the 4a to 4c represent CD errors, if only in a single field point P0, namely in the middle of the field, a mask interpretation is performed. In the 8a to 8c are curves of CD errors shown when at each field point PU, ie on each finger of the intensity variation device, a mask design is performed. Clearly recognizable are those in the 8a to 8c compared to the 4a to 4c significantly reduced CD errors due to the greater computational effort.

As already mentioned above, one main reason for the large values of the CD errors at field points P between the field points PU is the variation of the illumination intensity and, to a certain extent, the illumination direction distribution across the width of each finger of the intensity variation device.

An intensity variation ΔI can be converted into a prediction for the CD variation ΔCD of a structure by means of the so-called normalized intensity logarithm squared (NILS) according to the following mathematical relationship:

wobei der Ausdruck an der Position x0, an dem die Linienkante gedruckt werden soll, berechnet werden muss. Der NILS-Wert für jede Struktur kann somit ohne Mehraufwand bei der Maskenauslegung berechnet werden, wobei dieses häufig bereits automatisch geschieht.The NILS can be calculated directly from the profile I (x) of the aerial image intensity of the structure concerned, according to the following mathematical relationship:

the expression has to be calculated at the position x0 where the line edge is to be printed. The NILS value for each structure can thus be calculated without additional overhead in the mask design, which often happens automatically.

Thus, by means of equation (1), an intensity profile, as described, for example, in FIG 3 is converted into a prediction for the CD error caused thereby. This prediction can then be used to adjust the mask so that the CD error is compensated as completely as possible.

The so-called MEEF (Mask Error Enhancement Factor) indicates how much the width of a structure printed on the wafer changes as the width b of a structure on the mask changes. For the CD variation applies: ΔCD = β × MEEF × Δb (3)

Other known definitions of the MEEF include the magnification β in the size of the MEEF. Also, the MEEF can be calculated during the mask design advantageously without additional effort. The values for NILS and MEEF are thus available for each structure separately at each field point PM on which a mask design has been performed. By combining the equations (1) and (3), a correction of the mask widths can be calculated, which compensates the effect of the local intensity variation:

If the correction is carried out in accordance with equation (4), the results achieved according to the invention, as shown in FIGS 5 to 7 and 9 to 13 are shown qualitatively.

Equation (4) describes a rule-based mask adaptation, which can advantageously be carried out for a correction of feature sizes on the mask without much effort. This procedure also advantageously requires no change in the known processes of the mask manufacturer.

Course of inventively reduced CD errors are in the 9a to 9c and 13a to 13c shown. Here are in the 9a to 9c Gradients of the CD error when performing the correction according to equation (4) in a mask design shown only in the middle of the field. In the 13a to 13c are curves of the CD error in a mask design at each field point P with a correction according to equation (4).

The correction method proposed according to the invention can not be better than the value of the CD errors which results at the field points PU of the fingers of the intensity variation device. This minimum CD error results from the location-dependent change in the illumination direction distribution.

If a mask design is carried out at each field point PU (see 8a to 8c and 13a to 13c ), this minimum CD error will be zero. The comparison of 13a to 13c with the 8a to 8c indicates that the CD error can be reduced by one order of magnitude by means of the correction of the feature sizes according to equation (4).

If, on the other hand, a mask design is performed only in the middle of the field, the minimum CD error is significantly greater. Even if the reduction of the high-frequency contribution to the CD error is similarly good, ie that the absolute difference of the CD errors of the 4a to 4c and 9a to 9c is about the same as the difference of the CD errors of 8a to 8c and 13a to 13c , the relative reduction of the total CD error is only a factor of 2. This can be explained by the fact that the CD errors in 8th are very small, the required mask design disadvantageous but requires a very large computational effort and is therefore rarely used in practice.

The two options described so far were that a mask design is carried out either at each field point PU or only in the middle of the field P0. The variant of performing a mask interpretation at each field point PU can not always be practicable because of the necessary computational effort. If, on the other hand, the mask is only designed in the middle of the field by means of aerial image calculations, the CD error which can not be corrected by equation (4) due to the field dependence of the illumination direction distribution can be undesirably large depending on the application and the far field.

Therefore, as a further alternative, a third option is proposed, which provides that the mask design is carried out at three field points PM, for example in the middle of the field and +/- 2/3 of the Retikelhalbmessers or +/- 4/5 of Retikelhalbmessers. The feature size on the mask before then possibly a correction according to equation (4) is performed by a predefined interpolation (eg, linear, quadratic, parabolic, etc. interpolation) of the feature size on the mask at the two nearest determines three design field points PM or the three design field points PM. The use of only two design field points PM is also conceivable, just as the use of more than three design field points PM is possible.

The 5a to 5c and 10a to 10c show inventively reduced CD errors with ( 10a to 10c ) or without ( 5a to 5c ) Application of the corrective measure according to equation (4). Accordingly, it can be seen in particular from the figures that the CD error is equal to zero at the three selected mask design support points PM in the middle of the field and +/- 2/3 field width (ie at 0 and approximately +/- 35 mm).

Another alternative is to place each of two of the three mask design bases at the edge of the reticle area used. This choice is particularly useful if the course of the CD error is determined by the behavior at the edge. This case is in 6a to 6c and 11a to 11c shown. It is in the 6a to 6c for three different light sources A, B and C a mask design with subsequent linear interpolation between the field points PM shown. In the 11a to 11c For the three light sources A, B and C, a mask interpretation with linear interpolation and corrective action according to equation (4) is shown.

As another alternative are in the 7a to 7c and 12a to 12c still shows gradients of CD errors in a mask design in the middle of the field and in +/- 4/5 field width. It was in the 7a to 7c a quadratic interpolation is performed between said field points PM. In the 12c to 12c was also a correction of the structure widths according to equation (4) was performed.

In summary, a method for producing a mask for a lithographic illumination system according to the invention is proposed, with which high-frequency contributions to a CD error can be substantially completely removed by performing rule-based mask adjustments. Such rule-based mask adjustments are standard processes with mask manufacturers, whereby the rules proposed according to the invention for the correction according to equation (4) or the interpolation can advantageously be carried out without much additional effort.

As a result, as a residual effect for the CD error, a long-range, d. H. low-frequency curve with a typical period of about half the reticle width. This residual effect could be completely removed if a mask design were performed on a sufficient number of field points. However, this is often impractical from the computational effort. Therefore, as a modification of the invention it has been shown that with a mask design at only three field points with subsequent mask designation correction a significant improvement of the CD error can be achieved.

Advantageously, it is thus possible by means of the method according to the invention to produce improved masks with simple, low-computation-requiring rule-based correction methods which cooperate better with known illumination systems (eg EUV illumination systems) or better exploit their possibilities.

The person skilled in the art will suitably modify or combine the described features without departing from the essence of the invention.

Claims (9)

Method for producing a mask for a lithographic illumination system, comprising the following steps: a) setting an intensity variation device such that the most uniform possible illumination intensity is provided; b) determining a lighting direction distribution at a defined number of field points (PM) on the mask as a function of a first coordinate after averaging over a second coordinate; c) calculating feature sizes at the field points (PM) so that structures of the desired size are obtained in the resist pattern; and d) Creating the mask, wherein the feature size (b) on the mask is given at a location by a defined interpolation of feature sizes (b) on the mask at the field points (PM). A method for producing a mask for a lithographic illumination system, comprising the following steps: a) setting an intensity variation device in such a way that the illumination intensity that is as uniform as possible is provided; b) determining a lighting direction distribution at a defined number of field points (PM) on the mask as a function of a first coordinate after averaging over a second coordinate; c) calculating feature sizes at the field points (PM) so that structures of the desired size are obtained in the resist pattern; d) determining the illumination intensity at the largest possible number of field points; and e) creating the mask taking into account a correction of the feature sizes according to the following mathematical relationship:

with the parameters: Δb correction of the structure size CD desired structure width MEEF Mask-Error-Enhancement-Factor NILS Normalized-Intensity-Logarithm-Squared β Magnification scale ΔI (x) / I residual intensity error as a function of the x-coordinate A method for producing a mask for a lithographic illumination system, comprising: a) setting an intensity variation device in such a way that the illumination intensity as uniform as possible is provided; b) determining a lighting direction distribution at a defined number of field points (PM) on the mask as a function of a first coordinate after averaging over a second coordinate; c) calculating feature sizes at the field points (PM) so that structures of the desired size are obtained in the resist pattern; d) determining the illumination intensity at the largest possible number of field points (P); e) Defined interpolation of feature sizes (b) on the mask at the two field points (P) closest to each field point (P); and f) creating the mask taking into account a correction of the feature sizes according to the following mathematical relationship:

with the parameters: Δb correction of the structure size on the mask CD desired structure width MEEF Mask-Error-Enhancement-Factor NILS Normalized-Intensity-Logarithm-Squared β Magnification scale ΔI (x) / I residual intensity error as a function of the x-coordinate Method according to one of claims 1-3, wherein in step b) the determination of the illumination direction distribution at all field points (PU), at which an independent intensity adjustment is possible, is performed. Method according to one of claims 1-3, wherein in step b) the determination of the illumination direction distribution at a number of field points (PM) in the range between three and five is performed. Method according to one of the preceding claims, wherein in step d) between the field points (PM), the interpolation of the feature sizes (b) is one of: linear, square, parabolic. Method according to one of the preceding claims, wherein in step b) the determination of the illumination direction distribution at one or at all field points (PU), at which an independent intensity adjustment is possible, is performed. Method according to one of the preceding claims, wherein the field points (PM) are arranged in the middle of the mask +/- 2/3 Retikelhalbmesser. Method according to one of claims 1 to 7, wherein one of the field points (PM) is arranged at the edge of the mask.

Images